Within computing and information technology, interconnection systems are often modeled using graphs. A notable example is the swapped Optical Transpose Interconnection System (OTIS). Ensuring resilience to faults is crucial in optoelectronic systems. Across different fault classifications that can arise in an interconnection structure, two notable ones are the failure of a node (such as a processor in the case of OTIS over base graph G, denoted by OG and the disruption in facilitating communication between nodes breakdown of processor-to-processor communication). To manage these faults, it is crucial to assign a unique identifier to each node. In terms of graph theory, this corresponds to identifying the metric dimension, denoted as beta(G), and Fault-oriented metric dimension beta '(G) of the graph G representing the interlinking structure. This paper analyzes the OTIS architecture based on Cm (Cycle graph containing m vertices), represented as OCm, for resolvability and resolvability under fault conditions. It is proved that beta(OCm)=m-1 and beta '(OCm)=m.
A quasi total double Roman dominating function (QTDRD-function) on a graph G=(V(G),E(G)) is a function f:V(G)->{0,1,2,3} having the property that (i) if f(v)=0, then the vertex v must have at least two neighbors assigned 2 under f or one neighbor w with f(w)=3, and if f(v)=1, then the vertex v must have at least one neighbor w with f(w)>= 2, and (ii) if x is an isolated vertex in the subgraph of G induced by the set of vertices assigned nonzero values, then f(x)=2. The weight of a QTDRD-function f is the sum of its function values over the whole vertices, and the quasi total double Roman domination number gamma qtdR(G) equals the minimum weight of a QTDRD-function on G. The gamma qtdR-stability st gamma qtdR(G) (resp. gamma qtdR--stability st gamma qtdR-(G), gamma qtdR+-stability st gamma qtdR+(G)) of G is defined as the order of the smallest set of vertices whose removal changes (resp. decreases, increases) the quasi total double Roman domination number. In this paper, we first determine the exact values of gamma qtdR-stability for some special classes of graphs, and then we present some bounds on st gamma qtdR(G) and st gamma qtdR-(G). We also characterize the graphs with large st gamma qtdR-(G). Moreover, we show that if T is a nontrivial tree, then st gamma qtdR(T)<= 2, and if further T has maximum degree Delta >= 3, then st gamma qtdR-(T)<=Delta-1.
This paper investigates methods for constructing graphs whose matching polynomials are irreducible over Q. Building on this, the construction method is extended to general graph polynomials, for graphs whose polynomials satisfy certain properties, the connection construction produces a series of irreducible polynomial, such as characteristic polynomials and matching polynomials. Finally, the paper presents algorithms for calculating the matching polynomial of a given graph and the matching polynomial for a given number of vertices, implemented in Python.
Let T be a graph obtained by taking the Cartesian product of n cycles of even lengths >= 4. It is known that for any subset of k vertices of T with 1 <= k <= 2n-1, there exists a 2-factor in T where each cycle contains exactly one of the k vertices. We generalize this result to r-factors by proving that for 1 <= k <= 2n-r+1 and 2 <= r <= 2n, there exists an r-factor H in T such that each component of H is an r-regular, r-connected and bipancyclic graph, with each component containing exactly one of the k specified vertices of T. As a consequence, we obtain an analogous result for hypercubes.
In order to study the fault-tolerant capability of the multiprocessor system, several new self-diagnosis fault-tolerant models are proposed. In particular, non-inclusive diagnosability was showed by Ding et.al. in 2020. The non-inclusive diagnosability of regular graphs with C-3-free was established by Xu et. al. in 2022. In this paper, the non-inclusive diagnosability t(N)(AN(n)) of alternating group network is determined under the PMC model and the MM* model.
[Abstract] This paper aims to discuss the accuracy analysis of image recognition system based on deep learning algorithm and infrared thermal sensing. This paper expounds the basic principle and implementation of deep learning algorithm, and verifies its effectiveness through a series of experimental tests. Combining the advantages of cloud computing platform, a novel cloud computing image recognition system is designed, which makes full use of the characteristics of infrared thermal image data to train and optimize deep learning algorithms. The results show that the infrared thermal sensor image recognition system based on deep learning has significant advantages in precision analysis, the recognition rate is higher than the traditional algorithm, and it shows superior robustness in a variety of environments. The accuracy analysis shows that the cloud computing technology can effectively improve the processing capacity of the system, and further improve the efficiency and accuracy of image recognition.
This study aims at the problems of vehicle path recognition and equipment parameter acquisition, and proposes a solution based on Internet of Things positioning and thermal infrared image recognition. To this end, this paper combines Internet of Things technology with thermal infrared image analysis, aiming to achieve the automatic recognition of vehicle paths and the accurate measurement of key equipment parameters. This paper uses Internet of Things positioning technology (such as GPS) to obtain the initial location information of vehicles; Then, through thermal infrared image recognition technology, thermal image analysis of vehicles and related equipment is conducted, feature parameters are extracted, and combined with pattern recognition algorithms, the automatic recognition of vehicle operation paths is achieved. The results show that the method proposed in this paper can effectively identify the vehicle running path, and the recognition accuracy is significantly improved especially in complex environments. Therefore, the technology based on Internet of Things positioning and thermal infrared image recognition has significant advantages in the automatic recognition of vehicle paths and the acquisition of equipment parameters, providing new technical means for intelligent traffic management and vehicle safety monitoring.
In terms of training load evaluation of athletes, traditional evaluation methods often rely on physiological indicators and subjective feelings, and it is difficult to fully reflect the true state of athletes. This paper explores the application of infrared thermal energy image based on optical sensing in the evaluation of exercise training load, and studies the construction of an effective exercise thermal energy control model to improve the training effect and safety of athletes. Then, aiming at human motion monitoring, the mathematical model of motion information is established, and the monitoring accuracy is analyzed. The research shows that the infrared thermal energy image can accurately reflect the change of muscle temperature during exercise, and can reflect the load state of athletes more timely than the traditional monitoring method. Multi-sensor information fusion technology significantly improves the evaluation accuracy of motion load intensity, and verifies the feasibility of the model in practical application.
With the development of sports science, the integration of motion capture technology and thermal radiation imaging has provided new means for skeletal movement analysis. These technologies can more accurately capture and analyze the heat distribution and bone dynamics during movement. This paper analyzes the optimization parameters of skeletal movements based on thermal radiation imaging technology and motion capture systems. By constructing a motion heat model, it achieves precise capture and analysis of movement parameters, thereby providing data support and optimization recommendations for sports training. The study constructs a geometric model of human bones and performs collision detection on the model. Through parametric representation and sparse localization decomposition methods, the modeling and reconstruction of skeletal movements are carried out, analyzing the relationship between movement parameters and thermal effects. Research shows that the combination of motion capture and thermal radiation imaging technology can effectively enhance the accuracy of skeletal movement models.
This article proposes a visualization system for the evolution of art genres based on graph neural networks and time series analysis, aiming to reveal the classification and evolution trends of art works through advanced computer technology. The study provides an overview of relevant work, points out the limitations of existing methods for analyzing art works, and emphasizes the potential of graph neural networks and time series analysis in processing complex datasets. In the art classification algorithm based on graph neural network, this paper constructs a dataset for art works and designs graph tasks to achieve efficient information processing. The study explores clustering analysis algorithms based on time series networks and uses time series prediction models to provide in-depth interpretation of data presentation of art schools. The experimental results show that this algorithm has significant advantages in dynamic genre recognition. We have researched, designed, and implemented a visual system for classifying and evolving art genres. The system includes a dynamic theme classification model, research on the evolution path of art genres, and testing the accuracy of genre recognition, effectively achieving visual display of art classification and evolution. The results show that the system exhibits high accuracy and operability in practical applications.
A graph is 1-planar if it admits a drawing in the plane such that each edge is crossed at most once. Let G be a bipartite 1-planar graph with bipartition sets X and Y. A 1-disk [Formula: see text] drawing of G is a 1-planar drawing such that all vertices of X lie on the boundary of [Formula: see text] and all vertices of Y and all edges of G locate in the interior of [Formula: see text], where [Formula: see text] is a disk on the plane. The concept was first proposed by Huang, Ouyang and Dong when they solved a conjecture about the edge density of bipartite 1-planar graphs. Additionally, they presented a problem of determining the maximum number of edges in a bipartite graph with a 1-disk [Formula: see text] drawing. In this paper, we solve this problem and prove that every bipartite graph G which has a 1-disk [Formula: see text] drawing has at most [Formula: see text] edges. Moreover, we demonstrate that this upper bound is tight, in the sense that there are infinitely many graphs for which this bound is attained exactly.
Numerous studies have investigated application-specific clustering strategies for Wireless Sensor Networks (WSNs) to enhance their lifespan. The first WSN clustering method is LEACH. Cluster heads are randomly selected among sensor nodes. Each sensor node selects CHs randomly, depending on its residual energy according to a uniform distribution. Energy-constrained CH functions are cycled to balance. Although LEACH offers benefits, it also allows low-energy nodes to become CHs with the same likelihood as high-energy ones. This may lead to CH failure, resulting in reduced deployment cost and quality. An energy-efficient connection index-based clustering strategy for WSNs is presented to enhance LEACH. The Connectivity Index (CI) and Energy Consumption Ratio determine the selection of CH. REC prefers nodes with more residual energy, whereas CI guarantees CHs have more neighbors. Multiple pathways to the base station are considered in terms of energy, latency, rate, and stability. Energy utilization, latency, and transmissions decrease, extending network lifespan.
Graphs are useful for analyzing the structure models in computer science, operations research, and sociology. Also, different types of graph products have several applications in modeling, including those found in network analysis, communication protocols, organizational structures, diverse fields, network analysis and even chemistry. In this article, we introduce the concept of dominant strong metric dimension of graphs (DSMD for short) as a generalization of strong metric dimension and also dominant metric dimension. We determine the DSMD for several families of graphs that include for instance Km,n, the fan graph F1,n, the wheel graph W1,n and the helm graph Hn. Additionally, we study this invariant under the join product, corona product of two graphs and the generalized join graphs. Moreover, this article investigates the concept of dominant strong metric dimensions for some algebraic graphs associated with rings, which is the comaximal graph. In fact, by using the results on DSMD of the generalized join graphs, the dominant strong metric dimension of the comaximal graph of the ring of integers modulo n is investigated. Through this exploration, besides considering various graph products, we aim to provide a comprehensive framework for analyzing commutative rings and their associated graphs, thereby advancing both theoretical knowledge and practical applications in diverse domains.
Cloud communication is a combination of distributed computing and parallel computing. Task scheduling is a major challenge in cloud communications due to the NP-completeness of cloud systems. To address this, various swarm intelligence-based approximation techniques have been developed. This paper proposes a novel method for efficient task scheduling with improved security in cloud computing. A Hybrid Convolutional Neural Network with Long Short-Term Memory (HCNN-LSTM) optimized using FABOA is proposed for task scheduling to maximize throughput and minimize make span. Additionally, an improved random bit-stuffing technique with a modified RSA algorithm ensures secure data transmission. A novel Hybrid Convolutional Neural Network with Long Short-Term Memory (HCNN-LSTM) algorithm which is optimized using FABOA is proposed, which complicates readability. While the introduction outlines general cloud computing challenges, it lacks a focused literature review that identifies specific gaps in existing work and clearly justifies the need for the proposed HCNN-LSTM-FABOA system. Finally, our proposed approach is simulated under a cloudlet simulator and the evaluation results are analyzed to determine its performance. In addition to this, the proposed approach is compared with various other task scheduling-based approaches for various performance metrics, namely, resource utilization, response time, as well as energy consumption.
With the continuous development of computer technology, the field of fashion design increasingly emphasizes intelligence and personalization. A research has proposed an intelligent color scheme generation tool for clothing design based on K-means clustering and color transfer algorithm, aiming to improve the color matching effect and meet the personal style needs of modern consumers by optimizing the color extraction and transfer process. This article analyzes the shortcomings of current clothing design color matching tools and points out the limitations of existing systems in color extraction and application. Simulation results showed that this method can effectively extract color features and has good feasibility. By applying color transfer algorithms to different images, simulation results demonstrate the advantages of this method in terms of color consistency and visual effects. This article designs and implements an intelligent color matching system for clothing design, introducing the overall process of the color matching system, the structure of the color transfer module, and the establishment of the color matching model. After a series of tests, this tool has demonstrated efficient color matching capabilities in practical applications, meeting the diverse needs of designers for color matching.
Maize cultivation is frequently affected by diseases, and if they are not promptly prevented and managed, they can lead to reduced yield and quality, resulting in economic losses for farmers. Machine vision technology has been used to identify maize leaf diseases; however, detecting the Region of Interest (RoI) is challenging due to the irregular shapes of the affected areas. Otsu’s threshold-based segmentation offers precise RoI detection. Exhaustive threshold selection significantly degrades performance as the number of thresholds increases, causing exponential computational complexity. To address this, Averaging Histogram Equalization (AVHEQ) is used to denoise disease-affected images, and a novel Crossover Tuna Swarm Optimization (CTSO) algorithm determines the optimal segmentation threshold. Otsu’s method, a threshold-based approach, is also used for accurate RoI detection. In this study, AHEQ is applied to eliminate noise from disease-affected images, while a novel CTSO algorithm is employed to determine optimal segmentation thresholds of objective functions. The outcomes of ten-fold cross-validation were compared with those of previous studies utilizing assessment metrics like Accuracy (Acc), Precision (Pre), Sensitivity (Sen), and Error (Err).
Many patients are affected by Epilepsy, a prevalent neurological disorder. They are characterised by aberrant signal discharges on EEG (electroencephalogram). Visual examination of EEG recordings is labor-intensive, subjective, and needs substantial improvement as a non-invasive and low-cost method of identifying epileptic seizures. Hence, it is important to identify aberrant signals from large EEG data automatically, a task where CNNs (Convolution Neural Networks) and DL (deep learning) methods have received much attention. They face problems due to the large dimensionalities of data during training. This research work aims to examine the impacts of dimensionality reductions, feature extractions, and EEG signal identifications for seizure detections. BEMD (Bi-dimensional Empirical Mode Decomposition) has been introduced in this work for dimensionality reductions where features based on statistics, frequencies, and nonlinearity are extracted from sub-bands to collect adequate information on EEG signals. The LSTM (Long-Short Term Memory) model classifies EEG signals. In addition, to improve classification accuracy, BEMD reduces data dimensions resulting in reduced feature space. This work's proposed algorithm was evaluated on CHBMIT Scalp EEG Database in terms of Accuracy, Specificity, Sensitivity, and F1 scores where outcomes showed better performances when compared to other methods for automatic seizure detections.
Sometimes while you are using the Internet, for example, via a Wi-Fi network from one of the companies, the Internet is suddenly cut off due to a malfunction at that point, which disrupts your important work on the Internet, so there is a need for another source close to this point through which you can operate the Internet until this malfunction is fixed. To contribute to solving this problem, we assume that the two Wi-Fi points are D1 and D2, and the person using it is v such that v is adjacent to a vertex x is an element of D1 and a vertex y is an element of D2. In this paper, we introduce the concept of an alternative domination that models the aforementioned problem. Specifically, a dominating set D=D1 boolean OR D2 of a graph G=(V,E) is said to be an alternative domination if every vertex v is an element of V-D has at least one neighbor in D1 and one neighbor in D2. The cardinality of a minimum alternative dominating set in G is called the alternative domination number of G and is denoted by gamma 1,2(G). Basic properties and some interesting results have been obtained.
With the rapid development of computer vision and artificial intelligence, graph convolutional networks (GCNs) have shown great potential in processing noneuclidean data. In this study we review relevant work, explore existing sculpture image classification techniques and their limitations, and subsequently propose a novel sculpture image classification algorithm based on graph convolutional networks. The algorithm designed a graph convolutional network model that gradually updates node features through convolutional layers to effectively classify image content. The results indicate that the classification algorithm significantly outperforms traditional methods in terms of accuracy. The modeling and generation methods for 3D sculpture creation space include optimized 3D modeling processes and the application of hyperbolic space, implementing internal and external view extraction modules to improve the model’s ability to capture complex sculpture shapes. This paper proposes a new similarity measurement method for 3D sculpture models, which calculates the shape features of sculptures and analyzes the impact of different features on similarity evaluation through ablation experiments. The research results indicate that this method has good performance in identifying and classifying 3D sculpture styles, and provides theoretical support for future art creation and work analysis.
In this paper, we give some classes of Brandt semigroups which can be uniquely determined by their power graphs. Additionally, it is proved that the completely 0-simple semigroup whose automorphism group is same as that of its power graph is precisely K40 .
